compaction

compaction is an agent for coding agents from iamironz/ai-config-bundle. It costs 27 tokens per session (80 once invoked), scanned A, original, MIT.

A memory-preservation workflow for long coding tasks. It compresses important work state, unresolved questions, evidence, and next steps so the task can continue after context is shortened.

In plain words
What is it for?
Use it to preserve progress, open questions, delegated work, checkpoints, and evidence during extended tasks.
Why use it?
Long conversations can lose details when their context is compressed. This keeps the important decisions and unfinished work available.

Agent

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/iamironz/ai-config-bundle/compaction
Clone the repo
git clone --depth 1 https://github.com/iamironz/ai-config-bundle

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for compaction

README.md
[![agentmods](https://agentmods.dev/badge/agents/iamironz/ai-config-bundle/compaction.svg)](https://agentmods.dev/agents/iamironz/ai-config-bundle/compaction)
Your own site
<a href="https://agentmods.dev/agents/iamironz/ai-config-bundle/compaction"><img src="https://agentmods.dev/badge/agents/iamironz/ai-config-bundle/compaction.svg" alt="Measured on agentmods" height="20"></a>
Per session 27 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 80 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00027 $0.00080
Opus 5 $0.00014 $0.00040
Sonnet 5 $0.00005 $0.00016
Haiku 4.5 $0.00003 $0.00008

Measured 3d ago against content hash 64bdf9d2daa7, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

compaction scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

ai-kb/agents/compaction.md · 7 lines

What it actually says

Use this lane only to preserve continuity across compaction. Keep unresolved tasks, delegation state, checkpoints, overlays, critical evidence, and next actions intact. Do not start new implementation work or ask the user to continue.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 3d ago First seen · 7 lines · 27 tokens per session scan A 64bdf9d2daa7

Subscribe to this mod's changes

compaction is an agent published in the GitHub repository iamironz/ai-config-bundle (2 stars, last pushed 5mo ago), licensed MIT. It adds 27 tokens to every session and 80 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other agents, from other repositories

MEMORY

Generalized reusable lessons from agent sessions. Root causes converted into preventive rules, not incident-specific notes. Entries are h3 headers with [ACTIVE|RETIRED] status. Content: brief, grep-friendly, MECE across sections. Style: one-liner per entry, optional sub-bullets for context.

griddynamics/rosetta · 0 tokens

kmd-intake

The source-intake agent — drains the un-ingested sources queue. Delegate to it when raw material has been dropped into the KB's sources/ directory (scraped pages, whitepapers, transcripts, meeting notes) and needs to be distilled into KB pages, when lint reports unreferenced sources, or on a scheduled intake run.…

yasik/kmd · 84 tokens

kmd-operator

The knowledge-base interface agent. Delegate to it for any KB work — answering questions from a markdown knowledge base (queries with citations), writing knowledge into it (ingesting sources, promoting artifacts/reports, filing learnings, correcting pages), or deciding where something belongs in the KB. Works with…

yasik/kmd · 93 tokens

clippy-strict-validator

Use this agent when code changes have been made and need validation before completion or commit. Uses parallel background execution for 3-5x faster validation. Runs clippy, tests, and pattern scans simultaneously. The agent proactively ensures code quality standards are met.\n\nExamples:\n- User: "I've just added a…

dravr-ai/dravr-embacle · 0 tokens

error-handling-guardian

Guards structured error handling, preventing regression to anyhow and ensuring RunnerError usage.

dravr-ai/dravr-embacle · 21 tokens

copilot

GitHub Copilot reads .github/copilot-instructions.md automatically in VS Code and JetBrains IDEs. aimemo setup writes instructions into that file telling Copilot to maintain the memory loop: inject context at session start, log after file edits, recap at session end.

rustkit-ai/aimemo · 0 tokens